/ "-^ a ^ o ' ^} yjb ^ BULLETIN 806 SEPTEMBER 1982 4-3 Feasibility of Inactivating Phosphorus with Aluminum Salts in Ball Pond, CT By Wendell A. Norvel BALL POND NEW FAIRFIELD, CONN. TRACED FROM AERIAL SURVEY MAP 89 ACRES PLANIMETER MEASUREMENT CONTOUR INTERVAL 3 FEET ° 1 S ?" SCALE l"= 600' GOVERu;;,C:, , PUBLICATIONS RECEIVED NOV 23 198a UNIVERSITY LIBRARY UNIVERSITY OF CONHtCTiCUT Bathymetric map of Ball Pond. Connecticut State Board of Fisheries and Game, 1959. THE CONNECTICUT AGRICULTURAL EXPERIMENT STATION NEW HAVEN ,w s ii i Digitized by the Internet Archive in 2011 with funding from LYRASIS members and Sloan Foundation http://www.archive.org/details/feasibilityofinaOOnorv Feasibility of Inactivating Phosphorus with Aluminum Salts in Ball Pond, CT By Wendell A. Norvell The value of many lakes in Connecticut as aesthetic, recreational, and water-supply resources could be increased by reducing the availability of phosphorus (P) to algae (Norvell and Frink 1975; Norvell, Frink and Hill 1979). This bulletin evaluates the feasibility of reducing the availability of P in Ball Pond by adding soluble salts of aluminum (Al) to remove P from the v;ater and retard release of P from lake-bottom sediments. The treatment of lakes with Al to inactivate P is a method of lake restoration of relatively recent origin (Dunst et al. 1974; Jernelov 1971). This approach utilizes the well-known ability of Al to bind phosphate ions through precipitation and adsorption reactions which produce insoluble hydroxides of_ Al containing variable amounts of tightly bound P. Several of these treatments to inactivate P have been effective and all have provided valuable insight into the processes removing P from lakes, the equipment required for treatment, and the types of lakes that might benefit from treatment (Cooke and Kennedy 1981; Dunst et al. 1974; Peterson et al. 1973) . Ball Pond, located in New Fairfield, CT, was selected as a candidate for in-lake treatment with Al to inactivate P after considering the criteria listed below. These criteria for selection were modified and extended slightly from those suggested by Peterson et al. (1973): • The lake should be mesotrophic or eutrophic so that any improvement in lake condition is easily detectable and of significance to lake users. • The retention time of v/ater in the lake should be long enough to permit improvements to be observable and persistent. In general, this implies that the ratio of watershed to lake area should be relatively small. • Phosphorus should be the limiting nut- rient for planktonic algae or should effectively become the limiting nut- rient after treatment. •A substantial percentage of P in the lake should be in forms susceptible to inactivation at the time of treat- ment. • Inputs of P from the watershed should not be so large as to negate the effects of treatment. Phosphorus present in the lake or supplied to the lake from internal sources, e.g. sediments, should represent a substantial por- tion of the annual supply. •Lake depth should be great enough to per- mit effective settling and prevent resuspension of precipitated P from the bottom. • The surface area should be small enough to permit treatment at acceptable cost, however, if possible, large enough to allow generalization of results to other lakes. •The lake should have sufficient value for recreation or other uses to justify the expense of treatment. • Background data should be available or should be obtained on physical, chem- ical, and biological characteristics of the lake and its watershed. Connecticut Agricultural Experiment Station Bulletin 806 The characteristics of Ball Pond and its watershed meet many of the above criteria (Table 1, cover). Ball Pond is a small and moderately deep lake with a small ratio of watershed to lake area and a long water retention time of about five years. Considerable data on the lake and its watershed have been collected sporadi- cally since at least 1939. The lake is a recre- ational and aesthetic asset to nearby residents. It is stocked with trout by the Department of Environmental Protection and public access is provided at the State launching area at the southern end. The lake is eutrophic, v;ell sup- plied v;ith plant nutrients, and deficient in oxygen in the hypolimnion v;here hydrogen sulfide (HjS) is present. The ratio of total N to total P is usually greater than 20, v;hich is high enough so that P is expected to be the major nutrient limiting the populations of planktonic algae. These characteristics of Ball Pond and its watershed suggest that effective reduction of P concentrations v/ould significantly restrict the growth of algae, reduce the degree of eutro- phy, and increase the value of the lake as a recreational and fishery resource. Any benefits from inactivating P could persist for several years because of the long water retention time. Table 1. Characteristics of Ball Pond. Characteristic Value Reference Physical Surface area 36.4 ha 1,2 Vol ume 2.5 X 10^ m3 2 Maximum depth 15.9 m 2 Mean depth 6.9 m 2 Watershed/lake area ratio 2.6 1.5 Estimated water load 1.4 m/yr 5 Estimated retention time 5.0 yr 5 Land use Residential 39 % 5 Agricultural 15 % 5 Wooded 9 % 5 Lake surface 38 % 5 Chemical Total P, spring 35-40 ppb 3,5,6,7 Total N, spring 660-960 ppb 3,6,7 Alkalinity, spring 0.98 meq/1 7 N/P ratio, spring 16-35 3,6,7 Hypolimnetic H2S strong odor by mids umner 6 Hypolimnetic O2 depleted by midsunmer 1,2,6,7 Biological Secchi depth, spring 1.3-2.0 m 3,7 Secchi depth, summer 2.0-3.8 m 3,7 Chlorophyll-a, surimer Epilimnion 3 ppb 7 Metalimnion 25 ppb 7 Major planktonic algae Oscillatoria rubescens 3,7 Major rooted weeds Potamoqeton amplifolius 7 Ceratophyllum, sps 7 Elodea, sps 7 Weed beds, area! extent intermediate 7 Weed beds, density moderate 7 Connecticut State Board of Fisheries and Game (1942). Connecticut State Board of Fisheries and Game (1959). Connecticut Department of Environmental Protection (1979). Norvell and Frink (1975). Norvell, Frink, and Hill (1979). Norvell , unpublished Norvell (1980). The objectives of this investigation were to confirm the suitability of Ball Pond for treatments to inactivate P and to ansv;er spe- cific questions concerning the forms and amounts of P susceptible to inactivation, the proportion of P from external vs. internal sources, the choice of Al salts and their rate of applica- tion, the volume of the lake needing treatment, and whether or not such treatment could be made without damage to the fish population. METHODS Ball Pond was visited on April 11, July 1, July 24, August 26, October 15, and November 13, 1980 to collect water samples and measure dis- solved oxygen, temperature, and sulfide through- out the water column. Water samples were col- lected v;ith a 2 liter PVC Kemmerer sampler and stored in ice during transport to the labora- tory. Field analyses for dissolved 0^ and temper- ature were made with a YSI model 54RC oxygen- temperature meter and 30 meter probe. Field analyses for sulfide were made immediately after sample collection using a Bausch & Lomb Mini 20 spectrophotometer and sulfide reagent kit with a standard curve verified in the laboratory. Alkalinity was measured by titration using methyl purple as the end-point indicator. About one-half of the sample v;as filtered through a V7ell-rinsed 0.45 pm millipore filter. Soluble reactive P in the filtered portion was measured before freezing both the filtered and unfiltered sample. Later, water samples were thawed and analyzed for soluble P, total P, NH,,-N, and Kjeldahl-N using methods described by Norvell and Frink (1975). Nitrate-N was measured by a cadmium reduction me€hod (U.S. Environmental Protection Agency 1974), and SO,,-S v;as measured turbidimetrically as BaSO^ (American Public Health Association 1975). Aluminum was measured by atomic adsorption spectrophotometry following extraction by 8-hydroxyquinoline into methyl-i- sobutylketone (Amer. Publ. Health Assoc. 1975) and pH was determined potentiometrically . The relationship between area and depth in Ball Pond was determined by measuring the area between contours on the bathymetric map shown on the cover. This information V7as used to prepare a graph of the percentage of surface area below any depth (Fig. 1). This curve, in turn, was used to prepare the accompanying graph of the percentage of lake volume beneath any depth. These relationships and the area and volume data provided in Table 1 were combined with concen- tration data from lake samples to calculate the mass of each chemical constituent of interest within any depth interval of the lake. Mass balance calculations for P on different sampling dates were used to estimate internal release of Inactivating Phosphorus in Ball Pond P and the fraction of total P present as soluble reactive P, primarily inorganic orthophosphate . On July 24, 1980 a bulk sample of hypolim- netic water from the 12 meter depth was col- lected under oxygen-free conditions by displac- ing Ng from a glass carboy. The sample was packed in ice, kept in the dark, and prevented from contacting oxygen during transport to a laboratory cold room. One liter portions were treated under a Nj gas atmosphere with soluble Al in the form of diluted liquid alum following the general approach outlined by Cooke and Ken- nedy (1981). After mixing slowly for 5 minutes, each sample V7as allowed to stand for 15 minutes. The pH was then measured and the sample v;as fil- tered through a 0.45 )um pore size filter. One portion of the filtrate was titrated for alka- linity and another acidified for later analysis for soluble Al, SO^-S, soluble P, and soluble reactive P. 100 ^ 80 \ Ball Pond % Area and % Volume Below Depth 60 I \>- 40 Volume N^^ Areo 20 - ^^^^ 5 10 Depth, meters 15 Fig. 1 Percentage of surface area and volume of Ball Pond below any depth. RESULTS AND DISCUSSION Chemical Composition and Trophic Condition The results of field and laboratory analy- ses of water samples from Ball Pond are pre- sented in Tables 2 and 3. The data confirm that Ball Pond is indeed eutrophic. Concentrations of P and N were moderately high. The spring concentration of P was 37 ppb in 1980 which com- pares closely to concentrations of 35 ppb and 39 ppb that were found during spring overturn in 1977 and 1978 respectively. These concentra- tions are roughly two and one-half times the levels found in the 1939 Survey (Connecticut State Board of Fisheries and Game 1942), and are in the range of 20-50 ppb typical of eutrophic lakes. Spring N was also elevated at 0.89 ppm, and more than 20% of this N was present as read- ily available NO3". Phosphorus limitation of algae is strongly suggested by the spring N/P ratio of 24 and by the fact that free NO3" was present despite the on-going bloom of 0. rubes- cens . Another index of eutrophy is provided by hypolimnetic oxygen (Table 3) which v;as consumed rapidly during the spring and early summer. Between April 11, 1980 and July 1, 1980 virtu- ally all O2 in the hypolimnion v/as consumed, yielding a minimum rate of oxygen depletion of approximately 54 mg/cm^/day. This rate of hypo- limnetic oxygen depletion is in the middle of the 45-75 mg/cm^/day range typical of eutrophic lakes (Norvell and Frink 1975), and is roughly three times the rate found in 1939 (Connecticut State Board of Fisheries and Game 1942). In contrast to present conditions, oxygen was pres- ent in the upper hypolimnion even in August 1939. Chlorophyll-a concentrations (Norvell 1980) and algal numbers (Connecticut Department of Environmental Protection 1979) v;ere high during 1979 and 1980. The dominant algae in the lake during v/inter, spring, and most of the summer appears to be 0. rubescens. Algae counts taken from February through October 1979 indicated that this algae began "blooming" throughout the lake following fall overturn. As thermal stra- tification began in the spring, the algae con- centrated in the thermocline and remained there in declining numbers throughout the summer (Con- necticut Department of Environmental Protection 1979). In addition to planktonic algae. Ball Pond contains dense growths of Spatterdock (Nup- har sps.), Bigleaf Pondweed (Potomogeton sps.), Coontail (Ceratophyllum sps.), and filamentous algae (Spirogyra sps., Cladophora sps) in water less than two meters deep, particularly in the extensive shallow areas at the southern end (Norvell 1980) . During 1980, hydrogen sulfide was found throughout most of the hypolimnion, v;ith concen- trations approaching 2 ppm in the deeper waters. This too is an index of eutrophy, and also of major change since 1939 when Deevey noted that iron dominated the redox system rather than hydrogen sulfide (Connecticut State Board of Fisheries and Game 1942). At present, reducing conditions in the hypolimnion are much more severe with S0<,= undergoing reduction and HgS accumulating throughout the summer (Table 2, Fig. 2). The anoxic, sulfide-rich waters of the hypolimnion are not a suitable habitat for trout or most other organisms. These results confirm the eutrophy of Ball Pond and document the degradation in water qual- ity that has occurred during the last 40 years. Connecticut Agricultural Experiment Station Bulletin 806 Table 2. Chemical composition of water samples collected from Ball Pond during 1980. Date Depth Alkalinity Soluble Reactive P Soluble P Total P Organic Total N N meq/1 -ppb- 10-15 0 - 3.6 1.12 6 15 36 0.09 0.01 0.57 0.67 3.2 0.00 5 1.10 6 12 84 0.08 0.00 0.95 1.03 3.2 0.00 7 1.12 2 11 25 0.06 0.03 0.26 0.35 3.2 0.00 9 1.22 18 28 40 0.86 0.00 0.44 1.30 2.6 0.48 11 1.36 120 120 150 1.37 0.00 0.47 1.85 1.9 1.14 13 1.38 170 180 210 1.65 0.00 0.45 2.10 1.9 1.16 11 1.38 220 210 250 1.86 0.00 0.68 2.54 1.6 1.16 0 - 3.5 1.08 3 7 23 0.03 0.00 0.51. 0.54 3.2 0.00 5 1.04 2 8 40 0.01 0.00 0.62 0.63 2.9 0.00 7 1.08 2 15 42 0.21 0.01 0.46 0.68 2.6 0.00 9 1.30 28 34 66 0.90 0.00 0.47 1.37 2.2 0.38 11 1.40 160 160 200 1.69 0.01 0.71 2.41 1.3 1.40 13 1.43 230 240 250 2.09 0.00 0.75 2.76 1.0 1.65 H 1.82 250 250 340 2.20 0.00 0.65 2.85 0.6 1.90 0 - 4 0.93 1 6 17 0.12 0.00 0.45 0.57 3.0 0.00 5 1.08 1 4 20 0.04 0.00 0.51 0.55 3.2 0.00 7 1.08 1 6 38 0.05 0.00 0.62 0.67 2.6 0.00 9 1.36 11 14 63 1.10 0.00 0.95 2.05 1.1 0.39 11 1.50 190 200 230 2.32 0.00 0.94 3.26 0.7 1.40 13 1.48 260 270 280 2.88 0.00 0.48 3.36 0.8 1.80 14 1.48 280 290 310 2.79 0.00 0.77 3.56 0.6 1.80 0 - 6 0.88 0 13 15 0.05 0.00 0.34 0.39 3.0 0.00 7 1.16 1 4 16 0.03 0.00 0.50 0.53 2.9 0.00 8 1.20 1 4 36 0.31 0.00 0.63 0.94 2.6 0.35 9 1.58 8 14 70 1.08 0.01 0.86 1.94 3.2 0.68 11 1.68 172 162 207 2.18 0.00 0.77 2.95 1.0 1.70 13 1.94 310 306 354 2.96 0.00 1.31 4.27 0.3 1.90 14 1.91 268 270 294 3.09 0.01 0.73 3.81 0.6 1.90 11-13 0 - 14 Table 3. Temperature and dissolved oxygen data for Ball Pond during 1980. Depth, m Temperature ,'C 0 issolved O2, ppm 4-11 7-1 7-24 8-26 10-15 11-13 4-11 7-1 7-24 8-26 10-15 11-13 0 10.3 22.0 26.8 24.5 14.0 6.2 12.0 8.8 9.3 8.3 10.2 10.4 1 — — — 23.7 - - - - - 8.8 - - 2 - 21.0 26.6 23.3 13.2 6.1 - 9.0 9.1 9.0 9.2 9.6 3 - 21.8 26.0 22.7 - 6.1 - 9.2 9.0 9.4 - 9.5 4 7.8 16.0 19.2 22.1 13.2 6.0 11.8 11.2 9.4 9.0 8.5 9.4 5 - 13.0 15.0 18.4 13.2 6.0 - 1.4 7.6 8.6 8.7 9.4 6 - 10.0 11.2 13.8 13.2 6.0 - 0.6 2.0 4.2 9.1 9.3 7 - 9.0 9.7 10.2 13.2 6.0 - 0.3 0.5 0.5 9.0 9.3 8 7.0 - 8.7 8.9 10.0 6.0 11.6 - 0.3 0.3 0.6 9.2 9 - 8.0 8.1 8.1 8.4 6.0 - 0,2 0.3 0.2 0.4 9.2 10 -- -- 7.7 - 7.8 6.0 - - 0.2 - 0.4 9.1 11 - 7.3 7.1 7.2 7.2 6.0 - 0.2 0.2 0.2 0.3 9.1 12 5.8 - 7.0 - 7.1 6.0 10.5 - 0.2 - 0.3 9.1 13 - 7.0 7.0 7.1 7.0 6.0 - 0.2 0.2 0.2 0.3 9.2 14 5.8 7.0 6.9 7.0 7.0 6.0 10.2 0.2 0.2 0.2 0.3 9.1 15 -- 6.9 6.9 6.9 -- - -- 0.2 0.2 0.2 - -- Inactivating Phosphorus in Ball Pond Form, Distribution, and Internal Loading of P The success of P inactivation in lakes depends in part on the forms and distribution of P in the v;ater column. Table 2 shov;s the marked increase in P concentrations in the hypolimnion that occurred primarily during the early part of the summer. Decomposition of settling organic debris and release of sediment P under anoxic conditions both appear to have contributed to the elevated concentrations of P in the hypolim- nion. As a result of these processes, the mass of P in the lake is increased and redistributed into deeper waters during the summer. Redistri- bution is clearly illustrated by Fig. 3 which shows a pronounced peak in the mass of P in the middle of the hypolimnion at 11 meters. In con- trast to the results for April 11, 1980, the great majority of P v/as present in the hypolim- nion on August 26, 1980. In addition the total mass of P in the lake was higher by almost 35%. Of particular note in Table 2 is the pre- ponderance of soluble reactive P (largely or entirely orthophosphate) and the relatively small fraction of organic and particulate P in the hypolimnion during summer. A high propor- tion of soluble inorganic P is desirable for the objective of lake treatment because this form of P is readily precipitated by multivalent metals such as Al^"*" and is strongly adsorbed by many metal oxides and hydroxides, including Al (0H)3. The amounts of total and soluble reactive P in the lake are compared in Table 4 for six dates. E a. Q. I3> O. Q 3 Q. o .■s 8-26-80 20 - (Total 1251(g) 15 '\ 4-11-80 \ (Total 93 kg) A ■ 10 \s.,_^ y \ 5 '^^---Z^ ^^^^\- 0 1 1 1 1 1 1 L_ 15 Depth, m Fig. 2 Concentrations of sulfate and sulfide as a function of depth in Ball Pond. Fig. 3 Distribution of total P in Ball Pond on two dates. The mass of P contained in each one meter depth interval is shown as a function of depth. By late August, two- thirds of the total amount of P present at spring overturn v;as present in the hypolimnion as easily precipitated soluble reactive P. Nearly all of this v;as present below 10 meters, a region containing less than 11% of the volume of the lake. Relatively high concentrations of soluble reactive P in the hypolimnion persisted throughout the fall until autumnal circulation, when soluble reactive P virtually disappeared as oxygen mixed through the lake and the fall bloom of 0. rubescens began again. During the late summer and fall it is obvious that the form of P most susceptible to inactivation is concentrated in a region where desirable fish species would not be pres- ent. Results in Table 4 indicate that internal release of P contributes significantly to the summer increase in concentrations of P. Betv/een April 11, 1980 and July 1, 1980, total P increased by 52 kg or 56%. Work by Norvell, Frink, and Hill (1979) suggests that for Ball Pond and its watershed, the gross annual addi- tion of P to Ball Pond v;ould be about 73 kg with about 18 kg lost through the outflow. Thus, the net potential input is only 55 kg/year or about 4.6 kg/month. Even if none of this P v/ere lost by sedimentation or uptake by plants, external loading could account for only 12 kg of the 52 kg increase measured betv;een April 11, 1980 and July 1, 1980. The data suggest strongly that internal release of P constitutes a major input to the lake during the early summer. This con- tribution occurs during the same period as the onset of anoxia and severe reducing conditions. Release of P from the sediments certainly appears to be the most plausible source. Connecticut Agricultural Experiment Station Bulletin 806 Table 4. Distribution of total and soluble reactive P in Ball Pond on six dates in 1980. Region Depth Interval 4-11* 7-1 7-24 8-26 10-15 11-13 ( 0-15 m) Total P 93 Solb. Reac. P <13 Hypolimnion ( 7-15 m) Total P 28 Solb. Reac. P <4 Lower Hypolimnion (10-15 m) Total P 10 Solb. Reac. P <1 .45 139 125 121 116 53 65 64 64 3 64 88 90 94 33 44 61 62 62 1 46 59 63 67 12 39 49 57 57 <1 * Soluble reactive p was not measured on 4-11-8D. Values shown are for soluble P which includes both organic and inorganic P. The mass of soluble reactive P was probably 1/4 to 1/3 of these values. Table 5. Effect of alum additions on composition of hypolimnetic water from 12 m depth in Ball Pond. Chemicals Added Composition of Sample Liquid Alum Aluminum pH Alka- 1 inity Al SO4-S Soluble Reactive P Sol luble P Total P mg/1 mg/1 meq/1 meq/1 ppb ppm 0 0 0 7.11 1.34 0 1.0 208 216 249 30.7 1.3 0.14 6.80 1.16 12 2.6 6 9 - 61.4 2.6 0.29 6.68 1.00 5 - 3 6 - 123.0 5.2 0.58 6.50 0.67 12 10.2 2 6 - 184.0 7.8 0.87 6.20 0.46 20 15.0 2 9 - 245.0 10.4 1.16 5.70 0.28 52 18.9 3 6 - 307.0 13.0 1.44 5.23 0.08 610 22.7 3 9 - 368.0 15.6 1.73 4.65 0.00 >2500 26.9 3 3 _ Inactivating Phosphorus in Ball Pond During July through October, losses of P from the lake exceeded gains and the total mass of P present fell from 145 kg on July 1 to 125 kg on August 26, and to 121 kg on October 15, 1980. However, these losses occurred in the epilimnion and metalimnion only. As mentioned above, the mass of P found in the hypolimnion actually increased during this same period. Area, Volume, and Time of Treatment Treatment of Ball Pond with Al would have two specific objectives-. removal of P from the lake, and reduction in internal release of P following treatment. Efficient removal of P could be most easily achieved v;hen a major frac- tion was present in an easily precipitated form, when it was concentrated in a relatively small volume, and when the precipitated P could be deposited in deeper waters and removed from biological cycles. These conditions exist in the hypolimnion of Ball Pond during late July and August and persist through most of the fall. Fig. 3 and Table 4 show that both total P and soluble P were concentrated in the anoxic hypo- limnion below 7 meters, especially in the deep- est waters. The area and volume below 7 meters are 180,000 m^ (44.5 acres, 50% of lake surface area) and 723,000 m^ (191,000,000 gal., 29% of lake volume). Treatment of the hypolimnion would also achieve the second objective by cov- ering the sediments of the anoxic zone with a layer of flocculated A1(0H)3 to absorb any P released by the sediment before it could mix with the lake. Further, restricting the treat- ment to the anoxic hypolimnion v;ould have the important benefits of reducing any chance of deleterious effects on fish populations and reducing costs by decreasing the portion of the lake to be treated. Because soluble reactive P concentrated rapidly in the hypolimnion and persisted there throughout the summer, the timing of treatment does not appear critical. Successful treatment of the hypolimnion appears possible from late July through October. Treatment during Septem- ber or October would avoid disruption of recrea- tional uses during mid-summer. Source of Al and Rate of Addition Many salts of Al could serve as sources of soluble Al, e.g. alum, Al2(S04)3, sodium alumi- nate Na2Al204, and aluminum chloride AICI3 . Among the available sources, liquid alum appears to be the best choice for treatment of Ball Pond because of its lower cost, ease of application, and ready commercial availability (Cooke and Kennedy 1981; Loureiro Engineering Associates 1980). The amount of alum to be added to the hypo- limnion should be sufficient, at the very least, to precipitate, absorb, or entrap the great majority of P in a good settleable floe of Al hydroxide. When feasible, higher rates of addi- tion are desirable to cover the bottom sediments with a thicker layer of Al hydroxide to retard future releases of P. For this purpose. Cook and Kennedy (1981) suggest applying Al at the highest rate that will not depress pH to the point that residual dissolved Al could become toxic to fish. This limit is necessitated by the acidity released during hydrolysis of Al which lowers pH and retards further precipita- tion of Al hydroxide. A minimum pH of 6 and a maximum residual dissolved Al concentration of 50 iug/l appear to provide an adequate margin of safety, even for trout. The relations among P concentrations, pH, alkalinity, residual dissolved Al, and additions of alum are described below for hypolimnetic water taken from the 12 m depth in Ball Pond. Commercially available liquid alum was used as the source of Al. Experiments were conducted under an oxygen-free atmosphere and at a temper- ature of about 5 C to approximate conditions in the hypolimnion. Differing amounts of diluted liquid alum were added to portions of the water sample to give 0 to 15.6 mg/1 of added Al. The flocculated Al (0H)3, containing precipitated and absorbed P, was separated from soluble con- stituents by filtration. Alum additions were very effective in removing P as shown in Table 5 and Fig. 4. Even the lowest rate of addition, 1.3 mg/1 of Al, removed 96% of soluble P. Total P concentra- tions were reduced at least 83% simply by losses of soluble P. Actual removal of total P v7ould have been even somev;hat greater because of entrapment and settling of particulate matter. This small additional "benefit could not be meas- ured because of the experimental method used, but any losses of particulate matter would increase the overall effectiveness of treatment even further. The results suggest that almost any reasonable dose of alum would remove P effectively if well-mixed into the hypolimnion of Ball Pond. Additions of alum caused the pH to fall from 7.11 to 4.65 following a "buffer" curve typical of the weak acid HCO3" with a pK = 6.8 (Fig. 5). Alkalinity declined linearly with added alum until the pH decreased to close to 6. Below this pH, the hydrolysis of Al was less complete, alkalinity declined more slowly, and more Al remained soluble. An addition of 210 mg/1 of liquid alum or 8.9 mg/1 of Al was suffi- cient to depress the pH to 6. This addition of alum neutralized about 75% of the initial alka- linity of 1.34 meq/1. Residual dissolved Al Connecticut Agricultural Experiment Station Bulletin 806 200 3. O Q. 100 5 10 Al Added, mg/l Fig. 4 Concentrations of soluble reactive P and residual dissolved Al as functions of Al added to hypollmnetic v»ater from 12 m depth In Ball Pond. increased as the pH fell, especially below pH 5.5. Fig. 4 suggests that up to 10 mg/l of Al could have been added before residual dissolved Al rose above the 50 pg/1 limit suggested by Cooke and Kennedy (1981). This rate of alum addition was equivalent to 83% of the initial alkalinity. Thus, for control of both pH and dissolved Al, the maximum advisable rate for addition of alum would be equivalent to 75 to 83% of the initial alkalinity. Higher rates could be achieved with mixtures of acid-forming and alkali-forming salts, such as alum and sodium aluminate (Dominie 1980) , but that does not appear necessary for treatment of Ball Pond. A volume-weighted calculation of alkalinity indicates that about 940,000 equivalents were contained in the hypolimnion below a depth of 7 meters during the late summer. At an 80% neu- tralization rate, this alkalinity is equivalent to an addition of Al of about 6,800 kg or roughly 9.4 mg/l of Al in the treated volume. However, considering that the above rate is max- imal and that hydrolysis of Al will occur largely in the zone of application, a more con- servative rate of 5 mg/l or 3,700 kg of Al allows a greater margin of safety and should still be adequate for treatment. For the twin objectives of removing P from the water and covering sediments with a layer of Al hydroxide, the distribution of alum requires a compromise betv/een equal treatment per unit area and equal treatment per unit volume. For example, areas of the lake 7 to 11 meters deep could be treated at the 7 meter depth at the rate of 18 g/m^ of Al, v;hile areas greater than 11 meters in depth v/ould be treated at the 7 meter depth but at the rate of 24 g/m^ of Al. Such a treatment adds the desired total of about 3,700 kg Al to the hypolimnion but adds 1/3 more per unit area to the deeper sections of the hypolimnion. Other compromises could easily be designed and the ultimate choice may be strongly influenced by equipment capabilities. The only significant caveat with respect to applicatons of liquid alum arises from the SO4 anion associated with the soluble Al. Table 3 and Fig. 2 show that Ball Pond suffers at pres- ent from accumulation of HjS in much of the hypolimnion. Some hypolimnion HgS undoubtedly comes from the reduction of SO,, and major increases in concentrations of SO4 might enhance HgS accumulaton. Were this to occur, the higher HgS concentrations could be deleterious to trout and might even enhance release of P from sedi- ments. However, allowing for some loss of SO,, through lake outflow, anticipating some reduc- tion in hypollmnetic oxygen demand as a benefit of treatment, and by treating only the hypolim- nion at the conservative rate of 5 mg/l, the potential hazard from added SO4 appears small while the cost advantage of liquid alum is large (Loureiro Engineering Associates 1980). Inactivating Phosphorus in Ball Pond 300 Liquid Alum Added, mg/l Fig. S Change In pH, alkalinity, and added Al as a function of additions of liquid alum to hypollmnetlc water from 12 m depth of Ball Pond. CONCLUSIONS ACKNOWLEDGEMENTS Ball Pond is well-suited to treatment by aluminum salts for the inactivation of phospho- rus. The majority of P in the lake becomes con- centrated in the hypolimnion during summer stra- tification and a large proportion is present as easily precipitated inorganic phosphate. Labo- ratory studies suggest that treatment of the hypolimnion with liquid alum in late summer or early fall would remove about one-half of the total P in the lake, an amount somewhat greater than the estimated annual supply from external sources . Treatment of waters below 7 meters at a rate of 5 mg/l of Al should be adequate to remove P and provide a thin layer of aluminum hydroxide to retard future releases of sediment P. Such a treatment v;ould cover an area of 180,000 m2, adding 3,700 kg of Al to 730,000 m^ of hypolimnetic water. An application of about 85,000 kg of liquid alum would provide the desired addition of Al. Benefits from treatment should be relatively long lasting because of the large fraction of P removed and the five-year water retention time. The author thanks Jane Damschroder and Joseph Rapuano for technical assistance. LITERATURE CITED American Public Health Association. 1975. Standard Methods for the examination of v/ater and wastewater, 14th ed. Amer. Public Health Assoc. Washington, DC. 1193p. Connecticut Department of Environmental Protec- tion. 1979. Phase I Diagnostic Feasibility Study of Ball Pond. Hartford, CT. 95p. Connecticut State Board of Fisheries and Game. 1942. A fishery survey of important Connect- icut lakes. Bulletin No. 63. Hartford, CT. 339p. Connecticut State Board of Fisheries and Game. 1959. A fishery survey of the lakes and ponds of Connecticut. Report No. 1. Hart- ford, CT. 395p. 10 Connecticut Agricultural Experiment Station Bulletin 806 Cooke, G.D. and R.H. Kennedy. 1981. Precipita- tion and inactivation of phosphorus as a lake restoration technique. EPA-600/3-81-012 . U.S. Environ. Protection Agency. Corvallis, OR. Norvell, W.A. 1980. Trophic classification of Connecticut lakes. Final report to Connecti- cut Department of Environmental Protection. The Connecticut Agricultural Experiment Sta- tion. New Haven, CT. 55p. Dominie, D.R. 1980. Hypolimnetic aluminum treatment of softwater Annabessacook Lake. In Proceedings of EPA/OECD International Sym- posium on Inland Waters and Lake Restoration. EPA. In press. Norvell, W.A. and C.R. Frink. 1975. Water chemistry and fertility of twenty-three Con- necticut lakes. Bulletin 759. The Connecti- cut Agricultural Experiment Station. New Haven, CT. 45p. Dunst, R.C., et al. 1974. Survey of lake reha- bilitation techniques and experiences. Tech. Bull. No. 75. Dept. Natural Resources. Madi- son, WI . Norvell, W.A. , C.R. Frink, andD.E. Hill. 1979. Phosphorus in Connecticut lakes predicted by land use. Proc. Natl. Acad. Sci. USA. 76:5426-5429. Jernelov, A. 1971. Phosphate reduction in lakes by precipitation with aluminum sulfate. Proc. 5th Int. Water Poll. Res. Conf. Perga- mon Press Ltd. 1:1 15/1-15/16. Loureiro Engineering Associates. 1980. Engi- neering report on hypolimnetic phosphorus precipitation with alum at Ball Pond, New Fairfield, CT. Comm. No. 503-2, prepared for Connecticut Department Environ. Protec- tion. 25p. Peterson, J.O., J. P. Wall, T.L. Wirth, and S.M. Born. 1973. Eutrophication control: Nutrient inactivation by chemical precipitation at Horseshoe Lake, Wisconsin. Tech. Bull. No. 52. Dept. of Natural Resources. Madison, WI. 20p. U.S. Environmental Protection Agency. 1974. Methods for chemical analysis of water and wastes. EPA-625-16-74-003. Connecticut Libraries 39153028928291